|
1 """Translation helper functions.""" |
|
2 |
|
3 import locale |
|
4 import os |
|
5 import re |
|
6 import sys |
|
7 import warnings |
|
8 import gettext as gettext_module |
|
9 from cStringIO import StringIO |
|
10 |
|
11 from django.utils.importlib import import_module |
|
12 from django.utils.safestring import mark_safe, SafeData |
|
13 from django.utils.thread_support import currentThread |
|
14 |
|
15 # Translations are cached in a dictionary for every language+app tuple. |
|
16 # The active translations are stored by threadid to make them thread local. |
|
17 _translations = {} |
|
18 _active = {} |
|
19 |
|
20 # The default translation is based on the settings file. |
|
21 _default = None |
|
22 |
|
23 # This is a cache for normalized accept-header languages to prevent multiple |
|
24 # file lookups when checking the same locale on repeated requests. |
|
25 _accepted = {} |
|
26 |
|
27 # Format of Accept-Language header values. From RFC 2616, section 14.4 and 3.9. |
|
28 accept_language_re = re.compile(r''' |
|
29 ([A-Za-z]{1,8}(?:-[A-Za-z]{1,8})*|\*) # "en", "en-au", "x-y-z", "*" |
|
30 (?:;q=(0(?:\.\d{,3})?|1(?:.0{,3})?))? # Optional "q=1.00", "q=0.8" |
|
31 (?:\s*,\s*|$) # Multiple accepts per header. |
|
32 ''', re.VERBOSE) |
|
33 |
|
34 def to_locale(language, to_lower=False): |
|
35 """ |
|
36 Turns a language name (en-us) into a locale name (en_US). If 'to_lower' is |
|
37 True, the last component is lower-cased (en_us). |
|
38 """ |
|
39 p = language.find('-') |
|
40 if p >= 0: |
|
41 if to_lower: |
|
42 return language[:p].lower()+'_'+language[p+1:].lower() |
|
43 else: |
|
44 # Get correct locale for sr-latn |
|
45 if len(language[p+1:]) > 2: |
|
46 return language[:p].lower()+'_'+language[p+1].upper()+language[p+2:].lower() |
|
47 return language[:p].lower()+'_'+language[p+1:].upper() |
|
48 else: |
|
49 return language.lower() |
|
50 |
|
51 def to_language(locale): |
|
52 """Turns a locale name (en_US) into a language name (en-us).""" |
|
53 p = locale.find('_') |
|
54 if p >= 0: |
|
55 return locale[:p].lower()+'-'+locale[p+1:].lower() |
|
56 else: |
|
57 return locale.lower() |
|
58 |
|
59 class DjangoTranslation(gettext_module.GNUTranslations): |
|
60 """ |
|
61 This class sets up the GNUTranslations context with regard to output |
|
62 charset. Django uses a defined DEFAULT_CHARSET as the output charset on |
|
63 Python 2.4. |
|
64 """ |
|
65 def __init__(self, *args, **kw): |
|
66 from django.conf import settings |
|
67 gettext_module.GNUTranslations.__init__(self, *args, **kw) |
|
68 # Starting with Python 2.4, there's a function to define |
|
69 # the output charset. Before 2.4, the output charset is |
|
70 # identical with the translation file charset. |
|
71 try: |
|
72 self.set_output_charset('utf-8') |
|
73 except AttributeError: |
|
74 pass |
|
75 self.django_output_charset = 'utf-8' |
|
76 self.__language = '??' |
|
77 |
|
78 def merge(self, other): |
|
79 self._catalog.update(other._catalog) |
|
80 |
|
81 def set_language(self, language): |
|
82 self.__language = language |
|
83 |
|
84 def language(self): |
|
85 return self.__language |
|
86 |
|
87 def __repr__(self): |
|
88 return "<DjangoTranslation lang:%s>" % self.__language |
|
89 |
|
90 def translation(language): |
|
91 """ |
|
92 Returns a translation object. |
|
93 |
|
94 This translation object will be constructed out of multiple GNUTranslations |
|
95 objects by merging their catalogs. It will construct a object for the |
|
96 requested language and add a fallback to the default language, if it's |
|
97 different from the requested language. |
|
98 """ |
|
99 global _translations |
|
100 |
|
101 t = _translations.get(language, None) |
|
102 if t is not None: |
|
103 return t |
|
104 |
|
105 from django.conf import settings |
|
106 |
|
107 globalpath = os.path.join(os.path.dirname(sys.modules[settings.__module__].__file__), 'locale') |
|
108 |
|
109 if settings.SETTINGS_MODULE is not None: |
|
110 parts = settings.SETTINGS_MODULE.split('.') |
|
111 project = import_module(parts[0]) |
|
112 projectpath = os.path.join(os.path.dirname(project.__file__), 'locale') |
|
113 else: |
|
114 projectpath = None |
|
115 |
|
116 def _fetch(lang, fallback=None): |
|
117 |
|
118 global _translations |
|
119 |
|
120 loc = to_locale(lang) |
|
121 |
|
122 res = _translations.get(lang, None) |
|
123 if res is not None: |
|
124 return res |
|
125 |
|
126 def _translation(path): |
|
127 try: |
|
128 t = gettext_module.translation('django', path, [loc], DjangoTranslation) |
|
129 t.set_language(lang) |
|
130 return t |
|
131 except IOError, e: |
|
132 return None |
|
133 |
|
134 res = _translation(globalpath) |
|
135 |
|
136 # We want to ensure that, for example, "en-gb" and "en-us" don't share |
|
137 # the same translation object (thus, merging en-us with a local update |
|
138 # doesn't affect en-gb), even though they will both use the core "en" |
|
139 # translation. So we have to subvert Python's internal gettext caching. |
|
140 base_lang = lambda x: x.split('-', 1)[0] |
|
141 if base_lang(lang) in [base_lang(trans) for trans in _translations]: |
|
142 res._info = res._info.copy() |
|
143 res._catalog = res._catalog.copy() |
|
144 |
|
145 def _merge(path): |
|
146 t = _translation(path) |
|
147 if t is not None: |
|
148 if res is None: |
|
149 return t |
|
150 else: |
|
151 res.merge(t) |
|
152 return res |
|
153 |
|
154 for localepath in settings.LOCALE_PATHS: |
|
155 if os.path.isdir(localepath): |
|
156 res = _merge(localepath) |
|
157 |
|
158 for appname in settings.INSTALLED_APPS: |
|
159 app = import_module(appname) |
|
160 apppath = os.path.join(os.path.dirname(app.__file__), 'locale') |
|
161 |
|
162 if os.path.isdir(apppath): |
|
163 res = _merge(apppath) |
|
164 |
|
165 if projectpath and os.path.isdir(projectpath): |
|
166 res = _merge(projectpath) |
|
167 |
|
168 if res is None: |
|
169 if fallback is not None: |
|
170 res = fallback |
|
171 else: |
|
172 return gettext_module.NullTranslations() |
|
173 _translations[lang] = res |
|
174 return res |
|
175 |
|
176 default_translation = _fetch(settings.LANGUAGE_CODE) |
|
177 current_translation = _fetch(language, fallback=default_translation) |
|
178 |
|
179 return current_translation |
|
180 |
|
181 def activate(language): |
|
182 """ |
|
183 Fetches the translation object for a given tuple of application name and |
|
184 language and installs it as the current translation object for the current |
|
185 thread. |
|
186 """ |
|
187 if isinstance(language, basestring) and language == 'no': |
|
188 warnings.warn( |
|
189 "The use of the language code 'no' is deprecated. " |
|
190 "Please use the 'nb' translation instead.", |
|
191 PendingDeprecationWarning |
|
192 ) |
|
193 _active[currentThread()] = translation(language) |
|
194 |
|
195 def deactivate(): |
|
196 """ |
|
197 Deinstalls the currently active translation object so that further _ calls |
|
198 will resolve against the default translation object, again. |
|
199 """ |
|
200 global _active |
|
201 if currentThread() in _active: |
|
202 del _active[currentThread()] |
|
203 |
|
204 def deactivate_all(): |
|
205 """ |
|
206 Makes the active translation object a NullTranslations() instance. This is |
|
207 useful when we want delayed translations to appear as the original string |
|
208 for some reason. |
|
209 """ |
|
210 _active[currentThread()] = gettext_module.NullTranslations() |
|
211 |
|
212 def get_language(): |
|
213 """Returns the currently selected language.""" |
|
214 t = _active.get(currentThread(), None) |
|
215 if t is not None: |
|
216 try: |
|
217 return to_language(t.language()) |
|
218 except AttributeError: |
|
219 pass |
|
220 # If we don't have a real translation object, assume it's the default language. |
|
221 from django.conf import settings |
|
222 return settings.LANGUAGE_CODE |
|
223 |
|
224 def get_language_bidi(): |
|
225 """ |
|
226 Returns selected language's BiDi layout. |
|
227 |
|
228 * False = left-to-right layout |
|
229 * True = right-to-left layout |
|
230 """ |
|
231 from django.conf import settings |
|
232 |
|
233 base_lang = get_language().split('-')[0] |
|
234 return base_lang in settings.LANGUAGES_BIDI |
|
235 |
|
236 def catalog(): |
|
237 """ |
|
238 Returns the current active catalog for further processing. |
|
239 This can be used if you need to modify the catalog or want to access the |
|
240 whole message catalog instead of just translating one string. |
|
241 """ |
|
242 global _default, _active |
|
243 t = _active.get(currentThread(), None) |
|
244 if t is not None: |
|
245 return t |
|
246 if _default is None: |
|
247 from django.conf import settings |
|
248 _default = translation(settings.LANGUAGE_CODE) |
|
249 return _default |
|
250 |
|
251 def do_translate(message, translation_function): |
|
252 """ |
|
253 Translates 'message' using the given 'translation_function' name -- which |
|
254 will be either gettext or ugettext. It uses the current thread to find the |
|
255 translation object to use. If no current translation is activated, the |
|
256 message will be run through the default translation object. |
|
257 """ |
|
258 eol_message = message.replace('\r\n', '\n').replace('\r', '\n') |
|
259 global _default, _active |
|
260 t = _active.get(currentThread(), None) |
|
261 if t is not None: |
|
262 result = getattr(t, translation_function)(eol_message) |
|
263 else: |
|
264 if _default is None: |
|
265 from django.conf import settings |
|
266 _default = translation(settings.LANGUAGE_CODE) |
|
267 result = getattr(_default, translation_function)(eol_message) |
|
268 if isinstance(message, SafeData): |
|
269 return mark_safe(result) |
|
270 return result |
|
271 |
|
272 def gettext(message): |
|
273 return do_translate(message, 'gettext') |
|
274 |
|
275 def ugettext(message): |
|
276 return do_translate(message, 'ugettext') |
|
277 |
|
278 def gettext_noop(message): |
|
279 """ |
|
280 Marks strings for translation but doesn't translate them now. This can be |
|
281 used to store strings in global variables that should stay in the base |
|
282 language (because they might be used externally) and will be translated |
|
283 later. |
|
284 """ |
|
285 return message |
|
286 |
|
287 def do_ntranslate(singular, plural, number, translation_function): |
|
288 global _default, _active |
|
289 |
|
290 t = _active.get(currentThread(), None) |
|
291 if t is not None: |
|
292 return getattr(t, translation_function)(singular, plural, number) |
|
293 if _default is None: |
|
294 from django.conf import settings |
|
295 _default = translation(settings.LANGUAGE_CODE) |
|
296 return getattr(_default, translation_function)(singular, plural, number) |
|
297 |
|
298 def ngettext(singular, plural, number): |
|
299 """ |
|
300 Returns a UTF-8 bytestring of the translation of either the singular or |
|
301 plural, based on the number. |
|
302 """ |
|
303 return do_ntranslate(singular, plural, number, 'ngettext') |
|
304 |
|
305 def ungettext(singular, plural, number): |
|
306 """ |
|
307 Returns a unicode strings of the translation of either the singular or |
|
308 plural, based on the number. |
|
309 """ |
|
310 return do_ntranslate(singular, plural, number, 'ungettext') |
|
311 |
|
312 def check_for_language(lang_code): |
|
313 """ |
|
314 Checks whether there is a global language file for the given language |
|
315 code. This is used to decide whether a user-provided language is |
|
316 available. This is only used for language codes from either the cookies or |
|
317 session. |
|
318 """ |
|
319 from django.conf import settings |
|
320 globalpath = os.path.join(os.path.dirname(sys.modules[settings.__module__].__file__), 'locale') |
|
321 if gettext_module.find('django', globalpath, [to_locale(lang_code)]) is not None: |
|
322 return True |
|
323 else: |
|
324 return False |
|
325 |
|
326 def get_language_from_request(request): |
|
327 """ |
|
328 Analyzes the request to find what language the user wants the system to |
|
329 show. Only languages listed in settings.LANGUAGES are taken into account. |
|
330 If the user requests a sublanguage where we have a main language, we send |
|
331 out the main language. |
|
332 """ |
|
333 global _accepted |
|
334 from django.conf import settings |
|
335 globalpath = os.path.join(os.path.dirname(sys.modules[settings.__module__].__file__), 'locale') |
|
336 supported = dict(settings.LANGUAGES) |
|
337 |
|
338 if hasattr(request, 'session'): |
|
339 lang_code = request.session.get('django_language', None) |
|
340 if lang_code in supported and lang_code is not None and check_for_language(lang_code): |
|
341 return lang_code |
|
342 |
|
343 lang_code = request.COOKIES.get(settings.LANGUAGE_COOKIE_NAME) |
|
344 |
|
345 if lang_code and lang_code not in supported: |
|
346 lang_code = lang_code.split('-')[0] # e.g. if fr-ca is not supported fallback to fr |
|
347 |
|
348 if lang_code and lang_code in supported and check_for_language(lang_code): |
|
349 return lang_code |
|
350 |
|
351 accept = request.META.get('HTTP_ACCEPT_LANGUAGE', '') |
|
352 for accept_lang, unused in parse_accept_lang_header(accept): |
|
353 if accept_lang == '*': |
|
354 break |
|
355 |
|
356 # We have a very restricted form for our language files (no encoding |
|
357 # specifier, since they all must be UTF-8 and only one possible |
|
358 # language each time. So we avoid the overhead of gettext.find() and |
|
359 # work out the MO file manually. |
|
360 |
|
361 # 'normalized' is the root name of the locale in POSIX format (which is |
|
362 # the format used for the directories holding the MO files). |
|
363 normalized = locale.locale_alias.get(to_locale(accept_lang, True)) |
|
364 if not normalized: |
|
365 continue |
|
366 # Remove the default encoding from locale_alias. |
|
367 normalized = normalized.split('.')[0] |
|
368 |
|
369 if normalized in _accepted: |
|
370 # We've seen this locale before and have an MO file for it, so no |
|
371 # need to check again. |
|
372 return _accepted[normalized] |
|
373 |
|
374 for lang, dirname in ((accept_lang, normalized), |
|
375 (accept_lang.split('-')[0], normalized.split('_')[0])): |
|
376 if lang.lower() not in supported: |
|
377 continue |
|
378 langfile = os.path.join(globalpath, dirname, 'LC_MESSAGES', |
|
379 'django.mo') |
|
380 if os.path.exists(langfile): |
|
381 _accepted[normalized] = lang |
|
382 return lang |
|
383 |
|
384 return settings.LANGUAGE_CODE |
|
385 |
|
386 dot_re = re.compile(r'\S') |
|
387 def blankout(src, char): |
|
388 """ |
|
389 Changes every non-whitespace character to the given char. |
|
390 Used in the templatize function. |
|
391 """ |
|
392 return dot_re.sub(char, src) |
|
393 |
|
394 inline_re = re.compile(r"""^\s*trans\s+((?:".*?")|(?:'.*?'))\s*""") |
|
395 block_re = re.compile(r"""^\s*blocktrans(?:\s+|$)""") |
|
396 endblock_re = re.compile(r"""^\s*endblocktrans$""") |
|
397 plural_re = re.compile(r"""^\s*plural$""") |
|
398 constant_re = re.compile(r"""_\(((?:".*?")|(?:'.*?'))\)""") |
|
399 |
|
400 def templatize(src): |
|
401 """ |
|
402 Turns a Django template into something that is understood by xgettext. It |
|
403 does so by translating the Django translation tags into standard gettext |
|
404 function invocations. |
|
405 """ |
|
406 from django.template import Lexer, TOKEN_TEXT, TOKEN_VAR, TOKEN_BLOCK |
|
407 out = StringIO() |
|
408 intrans = False |
|
409 inplural = False |
|
410 singular = [] |
|
411 plural = [] |
|
412 for t in Lexer(src, None).tokenize(): |
|
413 if intrans: |
|
414 if t.token_type == TOKEN_BLOCK: |
|
415 endbmatch = endblock_re.match(t.contents) |
|
416 pluralmatch = plural_re.match(t.contents) |
|
417 if endbmatch: |
|
418 if inplural: |
|
419 out.write(' ngettext(%r,%r,count) ' % (''.join(singular), ''.join(plural))) |
|
420 for part in singular: |
|
421 out.write(blankout(part, 'S')) |
|
422 for part in plural: |
|
423 out.write(blankout(part, 'P')) |
|
424 else: |
|
425 out.write(' gettext(%r) ' % ''.join(singular)) |
|
426 for part in singular: |
|
427 out.write(blankout(part, 'S')) |
|
428 intrans = False |
|
429 inplural = False |
|
430 singular = [] |
|
431 plural = [] |
|
432 elif pluralmatch: |
|
433 inplural = True |
|
434 else: |
|
435 raise SyntaxError("Translation blocks must not include other block tags: %s" % t.contents) |
|
436 elif t.token_type == TOKEN_VAR: |
|
437 if inplural: |
|
438 plural.append('%%(%s)s' % t.contents) |
|
439 else: |
|
440 singular.append('%%(%s)s' % t.contents) |
|
441 elif t.token_type == TOKEN_TEXT: |
|
442 if inplural: |
|
443 plural.append(t.contents) |
|
444 else: |
|
445 singular.append(t.contents) |
|
446 else: |
|
447 if t.token_type == TOKEN_BLOCK: |
|
448 imatch = inline_re.match(t.contents) |
|
449 bmatch = block_re.match(t.contents) |
|
450 cmatches = constant_re.findall(t.contents) |
|
451 if imatch: |
|
452 g = imatch.group(1) |
|
453 if g[0] == '"': g = g.strip('"') |
|
454 elif g[0] == "'": g = g.strip("'") |
|
455 out.write(' gettext(%r) ' % g) |
|
456 elif bmatch: |
|
457 for fmatch in constant_re.findall(t.contents): |
|
458 out.write(' _(%s) ' % fmatch) |
|
459 intrans = True |
|
460 inplural = False |
|
461 singular = [] |
|
462 plural = [] |
|
463 elif cmatches: |
|
464 for cmatch in cmatches: |
|
465 out.write(' _(%s) ' % cmatch) |
|
466 else: |
|
467 out.write(blankout(t.contents, 'B')) |
|
468 elif t.token_type == TOKEN_VAR: |
|
469 parts = t.contents.split('|') |
|
470 cmatch = constant_re.match(parts[0]) |
|
471 if cmatch: |
|
472 out.write(' _(%s) ' % cmatch.group(1)) |
|
473 for p in parts[1:]: |
|
474 if p.find(':_(') >= 0: |
|
475 out.write(' %s ' % p.split(':',1)[1]) |
|
476 else: |
|
477 out.write(blankout(p, 'F')) |
|
478 else: |
|
479 out.write(blankout(t.contents, 'X')) |
|
480 return out.getvalue() |
|
481 |
|
482 def parse_accept_lang_header(lang_string): |
|
483 """ |
|
484 Parses the lang_string, which is the body of an HTTP Accept-Language |
|
485 header, and returns a list of (lang, q-value), ordered by 'q' values. |
|
486 |
|
487 Any format errors in lang_string results in an empty list being returned. |
|
488 """ |
|
489 result = [] |
|
490 pieces = accept_language_re.split(lang_string) |
|
491 if pieces[-1]: |
|
492 return [] |
|
493 for i in range(0, len(pieces) - 1, 3): |
|
494 first, lang, priority = pieces[i : i + 3] |
|
495 if first: |
|
496 return [] |
|
497 priority = priority and float(priority) or 1.0 |
|
498 result.append((lang, priority)) |
|
499 result.sort(lambda x, y: -cmp(x[1], y[1])) |
|
500 return result |
|
501 |
|
502 # get_date_formats and get_partial_date_formats aren't used anymore by Django |
|
503 # and are kept for backward compatibility. |
|
504 # Note, it's also important to keep format names marked for translation. |
|
505 # For compatibility we still want to have formats on translation catalogs. |
|
506 # That makes template code like {{ my_date|date:_('DATE_FORMAT') }} still work |
|
507 def get_date_formats(): |
|
508 """ |
|
509 Checks whether translation files provide a translation for some technical |
|
510 message ID to store date and time formats. If it doesn't contain one, the |
|
511 formats provided in the settings will be used. |
|
512 """ |
|
513 warnings.warn( |
|
514 "'django.utils.translation.get_date_formats' is deprecated. " |
|
515 "Please update your code to use the new i18n aware formatting.", |
|
516 PendingDeprecationWarning |
|
517 ) |
|
518 from django.conf import settings |
|
519 date_format = ugettext('DATE_FORMAT') |
|
520 datetime_format = ugettext('DATETIME_FORMAT') |
|
521 time_format = ugettext('TIME_FORMAT') |
|
522 if date_format == 'DATE_FORMAT': |
|
523 date_format = settings.DATE_FORMAT |
|
524 if datetime_format == 'DATETIME_FORMAT': |
|
525 datetime_format = settings.DATETIME_FORMAT |
|
526 if time_format == 'TIME_FORMAT': |
|
527 time_format = settings.TIME_FORMAT |
|
528 return date_format, datetime_format, time_format |
|
529 |
|
530 def get_partial_date_formats(): |
|
531 """ |
|
532 Checks whether translation files provide a translation for some technical |
|
533 message ID to store partial date formats. If it doesn't contain one, the |
|
534 formats provided in the settings will be used. |
|
535 """ |
|
536 warnings.warn( |
|
537 "'django.utils.translation.get_partial_date_formats' is deprecated. " |
|
538 "Please update your code to use the new i18n aware formatting.", |
|
539 PendingDeprecationWarning |
|
540 ) |
|
541 from django.conf import settings |
|
542 year_month_format = ugettext('YEAR_MONTH_FORMAT') |
|
543 month_day_format = ugettext('MONTH_DAY_FORMAT') |
|
544 if year_month_format == 'YEAR_MONTH_FORMAT': |
|
545 year_month_format = settings.YEAR_MONTH_FORMAT |
|
546 if month_day_format == 'MONTH_DAY_FORMAT': |
|
547 month_day_format = settings.MONTH_DAY_FORMAT |
|
548 return year_month_format, month_day_format |
|
549 |